No these SB-E chips have a higher TDP and put off more heat.
Printable View
No these SB-E chips have a higher TDP and put off more heat.
Based on what the PureOC resident overclocker said regarding temperatures between 1155 processors and 2011 processors... example overclocking an i7-3960x vs an i7-2600k over 5.0Ghz is easier with the larger sandy bridge-e processor due to the larger IHS surface. I also approve to that based on physics laws about force, I believe heat transfer works the same way. Like if you are exerting the same force over a smaller area compared to a larger area, weight and penetration would be larger with the smaller area. The theory supports the idea as to why ivys are hotter due to a smaller die aside from the theory of that "cheap and faulty" thermal paste... Or it could be both..Quote:
Originally Posted by PP Mguire
You are forgetting a few things though. HT produces heat, SB-E has 2 more cores which takes more power and produces more heat. A 2500k vs a SB-E at 5GHz the 2500k would be cooler. I can prove it right now. I had my old 2500k at 5GHz under my H50 and was able to actually boot and load temps were nominal enough not to be worried about it. I'm sitting at 4.5GHz right now with .25v less voltage than my 2500k, 500MHz less, and my load temps are almost as high as the 2500k that was faster.
The theory is correct, but a 2500 or 2600k will run cooler despite surface area and heat displacement.
Ivy Bridge is hotter because of the 3D transistor technology. You pack about 10% increase per core performance it's going to run hotter despite silicon shrinkage. Not only that, the 3570k and 3770k are clocked higher and turbo higher.
hehe good point, im just relaying what the guy from PureOC said :)
not really concerned with 2011 processors that much since I doubt I can use one..
I see that people really have different views regarding why ivy processors are hotter. most theories seem to be right anyway, I also believe in what you say about the improved performance causing higher temperatures.
Anyway, I would really like to hear tips from a pro overclocker :)
I do have a lot of questions like how many hours do you really need to run Prime95 to call the overclock "stable" or like what's the right offset, voltage and so on.
I wasn't disagreeing with the guy, I was just saying how it doesn't really help in terms of cooling because you are adding 2 more cores and 2 VR. As for Ivy Bridge, the theory was the TIM instead of solder used. Thing is, SB uses TIM as well. It's simply the newer transistors. If they only die shrinked SB then it would be cooler but less performance. The IB-E CPUs will utilize the same tech and I bet they run hotter too.Quote:
Originally Posted by najiro
Clock variables are different for each setup. I personally don't even run stress tests anymore when I OC. I'll let Intelburntest run like 20 loops with maximum RAM and threads and if that works I'm good. Nothing will eat away at my CPU more than that for that long as it stands.
Intelburntest? Is that a program? I actually agree with you. The friend that taught my overclocking (by phone and it was a lengthy talk) told me to run prime95 for 2 hours to check if it had "error" and conclude that the overclock is "stable".. However, when I joined the OC talks at pureOC, most are saying it's not enough LOL.. Most are saying it should be ran half a day... trolling? I don't really know LOL. I think it's such a waste of electricity..
Is this is?
http://www.techspot.com/downloads/49...lburntest.html
Can you teach me how to use that? Just install and run?
There were cases that when I ran prime95, BSOD occurs... so maybe that what my friend is trying to assess..
I think it's so boring to leave your system running for a day only to check its stability. But I torture my system with many tests. I found so heavy OCCT! And of course a lot of gaming and 3dmarking lol. Actually I'm undervolting it....(don't need to oc)...it's running fine with no problems :)
Probably more die in the bigger chips.Quote:
Originally Posted by najiro
No some tards just feel that running the program for 24 hours fully maximizing your CPU will show stability. That I agree in sure, but completely unnecessary. I look at it this way, if you have a quad core or above, no program besides a select few will actually run your CPU at 100% for any given length of time. At least one that a normal person would use on a daily basis. So I figure, if you can sustain some loops on IBT, do a run or 2 of 3dmark, and game then your stable enough. If you have crucial important things that require such a stable overclock then you probably shouldn't be overclocking as it is.Quote:
Originally Posted by najiro
IBT doesn't need to be installed. Just run it. The GUI is simple and you can understand it when you open it.
I did this at my last job for system burn-in over a 24hr period or overnight depending on the customer. We would run Prime95, OCCT, play a H.264 video, and one other thing all at the same time. But, since we were a SI, we needed to make sure it was 100% stable before sending it out to customers.Quote:
Originally Posted by PP Mguire
@PP Mguire
Yeah I see your point. No game or software can really max out a sandy bridge quad-core.. I do run 3dmark 11 and 3dmark vantage after an OC. So I guess those would be enough. How can I know that its okay then? As long as the program finished the loop with no BSOD?
@CDs
For you guys perhaps the prime95 burn is important. After all, you guys are selling the stuff. But do you personally do the same with your PC?
With my personal PCs, I would do the same thing just to ensure that it's operational and fully functional. However, I wouldn't run it overnight or anything like that. Maybe like an hour or two tops.Quote:
Originally Posted by najiro
i see.. yeah i think an hour or 2 is enough.. half a day is really wasteful of energy..Quote:
Originally Posted by CDsDontBurn
never heard of intelburntest before, just prime95 and furmark for burn tests. thanks for sharing guys :)
I still wouldn't do that. Overnight if you leave it takes a chance of over heating unless your cooling is just that good. Even still, after running a test for 24 hours the system can still crash due to the OC. It's why I have decided not to do that kind of stuff anymore as the strain on the system is unnecessary. Not only that, but programs like Furmark also put even more of a load than a regular program so even if it crashes if it ran for a long time odds have it no game or other program is going to cause a fail.Quote:
Originally Posted by CDsDontBurn
Quote:
Originally Posted by PP Mguire
From experience, I've never seen my computer go over a certain mark (which, of course, changes given the voltage, etc.). I'd argue that if it's sitting pretty at, say, 60 degrees Celsius, then it's not going to shoot up to 100 overnight.